Paths: File paths (shared/, references/, ../ln-*) are relative to skills repo root. If not found at CWD, locate this SKILL.md directory and go up one level for repo root. If shared/ is missing, fetch files via WebFetch from https://raw.githubusercontent.com/levnikolaevich/claude-code-skills/master/{path}.

Task Rework Executor

Executes rework for a single task marked To Rework and hands it back for review.

Purpose & Scope

• Load full task, reviewer comments, and parent Story; understand requested changes.
• Apply fixes per feedback, keep KISS/YAGNI, and align with guides/Technical Approach.
• Update only this task: To Rework -> In Progress -> To Review; no other tasks touched.

Inputs

Resolution: Task Resolution Chain. Status filter: To Rework

Task Storage Mode

MANDATORY READ: Load shared/references/tools_config_guide.md, shared/references/storage_mode_detection.md, and shared/references/input_resolution_pattern.md

Extract: task_provider = Task Management → Provider (linear | file).

File Mode transitions: To Rework → In Progress → To Review

MANDATORY READ: Load shared/references/mcp_tool_preferences.md — ALWAYS use hex-line MCP for code files when available. No fallback to standard Read/Edit unless hex-line is down.

Workflow (concise)

1. Resolve taskId: Run Task Resolution Chain per guide (status filter: [To Rework]).
2. Load task: Read task (Linear: get_issue; File: Read task file), review notes, parent Story. 2b) Goal gate: MANDATORY READ: shared/references/goal_articulation_gate.md — State REAL GOAL of this rework (what was actually broken, not "apply feedback"). Combine with 5 Whys (shared/references/problem_solving.md) to ensure root cause is articulated alongside the rework goal. NOT THE GOAL: a superficial patch that addresses the symptom, not the cause.
3. Plan fixes: Map each comment to an action; confirm no new scope added.
4. Implement: MANDATORY READ: shared/references/code_efficiency_criterion.md — Follow task plan/checkboxes; address config/hardcoded issues; update docs/tests noted in Affected Components and Existing Code Impact. Before handoff, verify 3 efficiency self-checks.
5. Quality: Run typecheck/lint (or project equivalents); ensure fixes reflect guides/manuals/ADRs/research.
6. Root Cause Analysis: Ask "Why did the agent produce incorrect code?" Classify: missing context | wrong pattern | unclear AC | gap in docs/templates. If doc/template gap found → update the relevant file (guide, template, CLAUDE.md) to prevent recurrence.
7. Handoff: Set task to To Review (Linear: update_issue; File: Edit status line); move it in kanban; add summary comment referencing resolved feedback + root cause classification.

Critical Rules

• Single-task only; never bulk update.
• Do not mark Done; only To Review (the reviewer decides Done).
• Keep language (EN/RU) consistent with task.
• No new tests/tasks created here; only update existing tests if impacted.
• Do NOT commit. Leave all changes uncommitted — the reviewer reviews and commits.

Definition of Done

• Task and review feedback fully read; actions mapped.
• Fixes applied; docs/tests updated as required.
• Quality checks passed (typecheck/lint or project standards).
• Root cause classified (missing context | wrong pattern | unclear AC | doc gap); doc/template updated if gap found.
• Status set to To Review; kanban updated; summary comment with fixed items + root cause.

Reference Files

• Tools config: shared/references/tools_config_guide.md
• Storage mode operations: shared/references/storage_mode_detection.md
• [MANDATORY] Problem-solving approach: shared/references/problem_solving.md
• Kanban format: docs/tasks/kanban_board.md

Version: 3.0.0 Last Updated: 2025-12-23